サイト内検索
AI モデル 2119
- gemma-3-12b-it-int4-awq↗
gaunernst
This is the QAT INT4 Flax checkpoint (from Kaggle) converted to HF+AWQ format for ease of use. AWQ was NOT used for quantization. You can find the conversion script `convert_flax.py` in this model repo. NOTE: this is NOT the same as the official QAT INT4 GGUFs released here https://huggingface.co/collections/google/gemma-3-qat-67ee61ccacbf2be4195c265b Below is the original Model card from https://huggingface.co/google/gemma-3-12b-it
- sdxl-turbo↗
Stability AI
SDXL-Turbo is a distilled version of SDXL 1.0, trained for real-time synthesis. SDXL-Turbo is based on a novel training method called Adversarial Diffusion Distillation (ADD) (see the technical report), which allows sampling large-scale foundational image diffusion models in 1 to 4 steps at high image quality. This approach uses score distillation to leverage large-scale off-the-shelf image diffusion models as a teacher signal and combines this with an adversarial loss to ensure high image fidelity even in the low-step regime of one or two sampling steps. - **Developed by:** Stability AI - **Funded by:** Stability AI - **Model type:** Generative text-to-image model - **Finetuned from model:** SDXL 1.0 Base For research purposes, we recommend our `generative-models` Github repository (https://github.com/Stability-AI/generative-models), which implements the most popular diffusion frameworks (both training and inference).
- DeepSeek-V4.1-Flash-1017-abliterated↗
mrfakename
This is deepseek-ai/DeepSeek-V4.1-Flash-1017 with its refusal behaviour removed by directional ablation. It keeps the same files, format and FP8 layout as the original, so it loads anywhere the original loads. The refusals in this model live in more than one direction: chat mode and think mode refuse along different directions. So instead of a single refusal direction, a rank-2 refusal **subspace** was removed: The strengths per layer came from a multi-objective search: minimise refusals and keep the first-token KL on harmless prompts low.
- Ornith-1.5-9B-uncensored-GGUF↗
ressl
**TL;DR:** abliterated Ornith-1.5-9B as a GGUF ladder (BF16, Q8_0, Q6_K, Q5_K_M, Q4_K_M + mmproj BF16); hard refusals reduced from 767/1120 to 20/1120 (NVFP4 sibling measurement) with the Q4_K_M build measured at 101/1120 on the same 5-dataset cross-eval. Same abliterated master as the NVFP4 repo, converted with llama.cpp for local and edge use. Built for security research, red-teaming and penetration-testing workflows; it will comply with requests a stock model refuses. Use responsibly. Effective refusal rate on 1120 harmful prompts across 5 datasets (temp 0, hard decline phrases in the first 25 words; naive keyword counts are inflated by compliant-with-disclaimer answers and are not the metric here). The NVFP4 column is the sibling repo's artifact, shown because the BF16 master it was measured on is identical; aggressive quantization visibly re-amplifies hedged declines (Q4 column).
- audioldm2↗
cvssp
AudioLDM 2 is a latent text-to-audio diffusion model capable of generating realistic audio samples given any text input. It is available in the 🧨 Diffusers library from v0.21.0 onwards. AudioLDM 2 was proposed in the paper AudioLDM 2: Learning Holistic Audio Generation with Self-supervised Pretraining by Haohe Liu et al. AudioLDM takes a text prompt as input and predicts the corresponding audio. It can generate text-conditional sound effects, human speech and music.
- resnet50_gn.a1h_in1k↗
timm
This model features: * ReLU activations * single layer 7x7 convolution with pooling * 1x1 convolution shortcut downsample Recipe details: * Based on ResNet Strikes Back `A1` recipe * LAMB optimizer * Stronger dropout, stochastic depth, and RandAugment than paper `A1` recipe * Cosine LR schedule with warmup - **Model Type:** Image classification / feature backbone - **Model Stats:** - Params (M): 25.6 - GMACs: 4.1 - Activations (M): 11.1 - Image size: train = 224 x 224, test = 288 x 288 - **Papers:** - ResNet strikes back: An improved training procedure in timm: https://arxiv.org/abs/2110.00476 - Deep Residual Learning for Image Recognition: https://arxiv.org/abs/1512.03385 - **Original:** https://github.com/huggingface/pytorch-image-models
- Qwen3.6-27B-GGUF↗
batiai
> 🆕 **Newer model available — Qwen3.8-27B GGUF** > Qwen's 2026-08 release. Same 27B class, six quants from 10 GB, vision included, Korean-verified. > `ollama run batiai/qwen3.8-27b:iq4` > **"Flagship Coding in a 27B Dense Package"** — imatrix-calibrated GGUF quantizations of **Qwen/Qwen3.6-27B** (dense, multimodal-capable) for on-device AI on Mac. > Built and verified by BatiAI for BatiFlow — free, unlimited, on-device AI automation. > 🔥 **100,000+ downloads** — the most-pulled model in the `batiai` org. Thanks for running it on your Mac. **Running this locally? BatiFlow turns it into a full on-device automation assistant — free.** > > _If it's useful, a ❤️ helps other Mac users find it._
- bert_uncased_L-4_H-128_A-2↗
Google
This is the set of 24 BERT models referenced in Well-Read Students Learn Better: On the Importance of Pre-training Compact Models (English only, uncased, trained with WordPiece masking). We have shown that the standard BERT recipe (including model architecture and training objective) is effective on a wide range of model sizes, beyond BERT-Base and BERT-Large. The smaller BERT models are intended for environments with restricted computational resources. They can be fine-tuned in the same manner as the original BERT models. However, they are most effective in the context of knowledge distillation, where the fine-tuning labels are produced by a larger and more accurate teacher. Our goal is to enable research in institutions with fewer computational resources and encourage the community to seek directions of innovation alternative to increasing model capacity.
- fullgreed↗
John2386
**Fullgreed** is a fine-tune of Alibaba Tongyi's **Z-Image** — the 6B-parameter Single-Stream Diffusion Transformer (S3-DiT) — tuned for photorealistic, phone-camera-authentic portraits and selfies: natural lighting, believable skin and hair texture, and outputs that read as real photos rather than "AI renders." It is a surgical fine-tune: only the attention and feed-forward projection weights were trained, leaving the base model's norms, embedders, and timestep conditioning untouched. Fullgreed keeps everything Z-Image is good at (bilingual prompt following, text rendering, composition) while adding its own photographic character. Both files produce the same images — the INT8 is effectively lossless (measured 0.03% average weight error vs the full-precision weights — visually identical output) at half the size and memory, and it needs **no custom nodes**: it runs on stock ComfyUI, including ComfyUI Cloud.
- wav2vec2-large-xlsr-53-italian↗
jonatasgrosman
Fine-tuned facebook/wav2vec2-large-xlsr-53 on Italian using the train and validation splits of Common Voice 6.1. When using this model, make sure that your speech input is sampled at 16kHz. This model has been fine-tuned thanks to the GPU credits generously given by the OVHcloud :) The script used for training can be found here: https://github.com/jonatasgrosman/wav2vec2-sprint